Grant: CCI Applied Research and Development Grants

Transmission pipelines are critical components of modern infrastructure and ensuring the safety and integrity of these pipelines is vital to prevent potential disasters and environmental hazards. Pipeline fences prevent unauthorized access to pipeline rights-of-way and minimize the risk of interference, accidents, or injuries. Above-Ground Markers AGMs typically contain information such as pipeline and company identification, safety warnings, and emergency contact. Effective detection and monitoring of markers, fences, and buildings near pipelines are essential to their safety, maintenance, and regulatory compliance. To enhance pipeline safety leveraging cutting-edge innovations in image processing and machine learning for identifying and monitoring such objects of interest is indispensable. Various deep-learning approaches may be used for object recognition in aerial imagery. The proposal presents a paradigm shift in pipeline monitoring, leveraging state-of-the-art technology and strategic partnerships to enhance safety, efficiency, and environmental sustainability. Through the fusion of aerial imagery and machine learning, we offer a proactive and holistic approach to safeguarding pipeline infrastructure and the communities it serves. The specific objectives of this project are: 1) Application of advanced image processing and deep learning methods for identification of changes such as movement, addition, or removal of markers, fences, and buildings along the pipeline routes. 2) Provide real-time information and notifications to stakeholders in the event of any detected changes or anomalies. Also, provides streamlined automated monitoring thereby optimizing operational efficiency. 3) Provide recommendations for data collection, data storage, and processing for operational efficiency of monitoring and intervention. 4) Design a scalable and adaptable monitoring framework ensuring the long-term effectiveness and relevance of the proposed solution. The advantages of the proposed solution are (1) a thorough deep learning approach to the detection of changes in the proximity of pipeline infrastructure, allowing for prompt reaction to changes, (2) the automated nature of our monitoring system reduces the reliance on manual inspections, thereby optimizing operational efficiency and reducing maintenance costs. Additionally, the use of aerial imagery ensures comprehensive coverage of vast and often challenging terrains, enabling seamless monitoring of pipelines.